X server - signal 4 (illegal instruction) loading glamoregl

stultumanto · 2024-07-24 21:05:41

Intel i7-4770MQ CPU + NVIDIA GeForce GTX 765 M GPU. I'm running Excalibur and just did an update today. X now fails to start, and I'm seeing the message below in the log. Searching turned up nothing helpful. I thought it might have something to do with the nouveau firmware I installed a few weeks ago, but removing the firmware made no difference.

I next thought it might be a problem with the nouveau driver, but looking at the log, the crash seemed to happen when loading the modesetting driver. At any rate, trying to disable the nouveau driver (by forcing the modesetting driver) had the exact same result.

The crash appears to happen specifically when glamoregl is loaded, so I tried forcing the server to load without glamoregl by putting the lines below into the file /etc/X11/xorg.conf.d/00-noglamoregl.conf. The log did initially say that glamoregl would not be loaded by default, however, glamorgl did load when the fbdev driver was loaded. At that point, it crashed just as before.

Section "Device"
	Identifier "nogpu"
	Driver "modesetting"
	Option "Accelmethod" "none"
EndSection

Section "Module"
	Disable "glamoregl"
EndSection

It appears glamoregl is the culprit, but I don't know how to disable it universally, or whether the server will even be usable without it. Any suggestions would be appreciated. At this point, I'm afraid I may just have to wait for an update. Below is the full text of the initial log.

[   165.321] 
X.Org X Server 1.21.1.11
X Protocol Version 11, Revision 0
[   165.321] Current Operating System: Linux alien 6.9.9-amd64 #1 SMP PREEMPT_DYNAMIC Debian 6.9.9-1 (2024-07-13) x86_64
[   165.322] Kernel command line: BOOT_IMAGE=/boot/vmlinuz-6.9.9-amd64 root=UUID=3eff8d67-d669-4093-816d-08f3ff918faa ro quiet
[   165.322] xorg-server 2:21.1.12-1devuan1 (https://www.devuan.org/os/community) 
[   165.322] Current version of pixman: 0.42.2
[   165.322] 	Before reporting problems, check http://wiki.x.org
	to make sure that you have the latest version.
[   165.322] Markers: (--) probed, (**) from config file, (==) default setting,
	(++) from command line, (!!) notice, (II) informational,
	(WW) warning, (EE) error, (NI) not implemented, (??) unknown.
[   165.322] (==) Log file: "/var/log/Xorg.0.log", Time: Wed Jul 24 12:31:21 2024
[   165.322] (==) Using config directory: "/etc/X11/xorg.conf.d"
[   165.322] (==) Using system config directory "/usr/share/X11/xorg.conf.d"
[   165.322] (**) Option "defaultserverlayout" "ServerLayout0"
[   165.322] (**) ServerLayout "ServerLayout0"
[   165.322] (==) No screen section available. Using defaults.
[   165.322] (**) |-->Screen "Default Screen Section" (0)
[   165.322] (**) |   |-->Monitor "<default monitor>"
[   165.322] (==) No monitor specified for screen "Default Screen Section".
	Using a default monitor configuration.
[   165.322] (**) Option "BlankTime" "0"
[   165.322] (**) Option "StandbyTime" "0"
[   165.322] (**) Option "SuspendTime" "0"
[   165.322] (**) Option "OffTime" "0"
[   165.322] (**) Allowing byte-swapped clients
[   165.322] (==) Automatically adding devices
[   165.322] (==) Automatically enabling devices
[   165.322] (==) Automatically adding GPU devices
[   165.322] (==) Automatically binding GPU devices
[   165.322] (==) Max clients allowed: 256, resource mask: 0x1fffff
[   165.322] (WW) The directory "/usr/share/fonts/X11/cyrillic" does not exist.
[   165.322] 	Entry deleted from font path.
[   165.322] (==) FontPath set to:
	/usr/share/fonts/X11/misc,
	/usr/share/fonts/X11/100dpi/:unscaled,
	/usr/share/fonts/X11/75dpi/:unscaled,
	/usr/share/fonts/X11/Type1,
	/usr/share/fonts/X11/100dpi,
	/usr/share/fonts/X11/75dpi,
	built-ins
[   165.322] (==) ModulePath set to "/usr/lib/xorg/modules"
[   165.322] (II) The server relies on udev to provide the list of input devices.
	If no devices become available, reconfigure udev or disable AutoAddDevices.
[   165.322] (II) Loader magic: 0x564b0d38eea0
[   165.322] (II) Module ABI versions:
[   165.322] 	X.Org ANSI C Emulation: 0.4
[   165.322] 	X.Org Video Driver: 25.2
[   165.322] 	X.Org XInput driver : 24.4
[   165.322] 	X.Org Server Extension : 10.0
[   165.322] (++) using VT number 7

[   165.322] (II) seat-libseat: libseat integration requires -keeptty and -keeptty was not provided, disabling libseat integration
[   165.322] (II) xfree86: Adding drm device (/dev/dri/card0)
[   165.322] (II) Platform probe for /sys/devices/pci0000:00/0000:00:01.0/0000:01:00.0/drm/card0
[   165.323] (II) xfree86: Adding drm device (/dev/dri/card1)
[   165.323] (II) Platform probe for /sys/devices/pci0000:00/0000:00:01.1/0000:07:00.0/drm/card1
[   165.325] (--) PCI:*(1@0:0:0) 10de:11e1:1028:05ab rev 161, Mem @ 0xd4000000/16777216, 0xa0000000/268435456, 0xb0000000/33554432, I/O @ 0x00005000/128, BIOS @ 0x????????/131072
[   165.325] (--) PCI: (7@0:0:0) 10de:11e1:1028:05ab rev 161, Mem @ 0xd2000000/16777216, 0xc0000000/268435456, 0xd0000000/33554432, I/O @ 0x00004000/128, BIOS @ 0x????????/524288
[   165.325] (II) LoadModule: "glx"
[   165.325] (II) Loading /usr/lib/xorg/modules/extensions/libglx.so
[   165.326] (II) Module glx: vendor="X.Org Foundation"
[   165.326] 	compiled for 1.21.1.11, module version = 1.0.0
[   165.326] 	ABI class: X.Org Server Extension, version 10.0
[   166.042] (==) Matched modesetting as autoconfigured driver 0
[   166.042] (==) Matched fbdev as autoconfigured driver 1
[   166.042] (==) Matched vesa as autoconfigured driver 2
[   166.042] (==) Assigned the driver to the xf86ConfigLayout
[   166.042] (II) LoadModule: "modesetting"
[   166.042] (II) Loading /usr/lib/xorg/modules/drivers/modesetting_drv.so
[   166.042] (II) Module modesetting: vendor="X.Org Foundation"
[   166.042] 	compiled for 1.21.1.11, module version = 1.21.1
[   166.042] 	Module class: X.Org Video Driver
[   166.042] 	ABI class: X.Org Video Driver, version 25.2
[   166.042] (II) LoadModule: "fbdev"
[   166.042] (II) Loading /usr/lib/xorg/modules/drivers/fbdev_drv.so
[   166.042] (II) Module fbdev: vendor="X.Org Foundation"
[   166.042] 	compiled for 1.21.1.3, module version = 0.5.0
[   166.042] 	Module class: X.Org Video Driver
[   166.042] 	ABI class: X.Org Video Driver, version 25.2
[   166.042] (II) LoadModule: "vesa"
[   166.042] (II) Loading /usr/lib/xorg/modules/drivers/vesa_drv.so
[   166.042] (II) Module vesa: vendor="X.Org Foundation"
[   166.042] 	compiled for 1.21.1.9, module version = 2.6.0
[   166.042] 	Module class: X.Org Video Driver
[   166.042] 	ABI class: X.Org Video Driver, version 25.2
[   166.042] (II) modesetting: Driver for Modesetting Kernel Drivers: kms
[   166.042] (II) FBDEV: driver for framebuffer: fbdev
[   166.042] (II) VESA: driver for VESA chipsets: vesa
[   166.042] (WW) xf86OpenConsole: setpgid failed: Operation not permitted
[   166.042] (WW) xf86OpenConsole: setsid failed: Operation not permitted
[   166.051] (II) modeset(0): using drv /dev/dri/card0
[   166.051] (WW) Falling back to old probe method for fbdev
[   166.051] (II) Loading sub module "fbdevhw"
[   166.051] (II) LoadModule: "fbdevhw"
[   166.051] (II) Loading /usr/lib/xorg/modules/libfbdevhw.so
[   166.051] (II) Module fbdevhw: vendor="X.Org Foundation"
[   166.051] 	compiled for 1.21.1.11, module version = 0.0.2
[   166.051] 	ABI class: X.Org Video Driver, version 25.2
[   166.051] (II) modeset(G0): using drv /dev/dri/card1
[   166.052] (II) modeset(0): Creating default Display subsection in Screen section
	"Default Screen Section" for depth/fbbpp 24/32
[   166.052] (==) modeset(0): Depth 24, (==) framebuffer bpp 32
[   166.052] (==) modeset(0): RGB weight 888
[   166.052] (==) modeset(0): Default visual is TrueColor
[   166.052] (II) Loading sub module "glamoregl"
[   166.052] (II) LoadModule: "glamoregl"
[   166.052] (II) Loading /usr/lib/xorg/modules/libglamoregl.so
[   166.055] (II) Module glamoregl: vendor="X.Org Foundation"
[   166.055] 	compiled for 1.21.1.11, module version = 1.0.1
[   166.055] 	ABI class: X.Org ANSI C Emulation, version 0.4
[   166.079] (EE) 
[   166.079] (EE) Backtrace:
[   166.079] (EE) 0: /usr/lib/xorg/Xorg (OsLookupColor+0x14d) [0x564b0d30243d]
[   166.080] (EE) 1: /lib/x86_64-linux-gnu/libc.so.6 (__sigaction+0x40) [0x7fce28b8e590]
[   166.080] (EE) 2: /usr/lib/x86_64-linux-gnu/dri/nouveau_dri.so (nouveau_drm_screen_create+0x1326) [0x7fce267cf4a6]
[   166.080] (EE) 3: ? (?+0x0) [0xb90]
[   166.080] (EE) 4: ? (?+0x0) [0x564b14bc4aa0]
[   166.080] (EE) 
[   166.080] (EE) Illegal instruction at address 0x7fce267cf4a6
[   166.080] (EE) 
Fatal server error:
[   166.080] (EE) Caught signal 4 (Illegal instruction). Server aborting
[   166.080] (EE) 
[   166.080] (EE) 
Please consult the The X.Org Foundation support 
	 at http://wiki.x.org
 for help. 
[   166.080] (EE) Please also check the log file at "/var/log/Xorg.0.log" for additional information.
[   166.080] (EE) 
[   166.090] (EE) Server terminated with error (1). Closing log file.

EDIT - I just noticed there's an error in there about seatd, I'm not sure why I would suddenly be having problems with that, or what to do about it. There's virtually nothing online about seatd. Someone else mentioned getting the same error on this forum:
https://dev1galaxy.org/viewtopic.php?id=6385
But the thread wasn't very helpful. I've had seatd installed and working on this system for many months.

Last edited by stultumanto (2024-07-24 21:51:11)

stargate-sg1-cheyenne-mtn · 2024-07-25 07:05:09

did you install/reinstall seatd as per ralph.ronnquist (and logout/login)? updates sometimes cause permission issues.

stultumanto · 2024-07-25 08:55:16

It seems the seatd issue is unrelated to the problem of X not starting on this system. The same message shows up on one of my other systems, where X starts successfully:

[    20.893] (II) seat-libseat: libseat integration requires -keeptty and -keeptty was not provided, disabling libseat integration

seatd has been installed on both systems for many months, and the seatd daemon is currently running on both. `dpkg-reconfigure seatd` didn't make the message go away. Removing and reinstalling seatd would require reinstalling the X server itself. I don't know whether this message represents a genuine problem on either system, but it doesn't seem to be related to the signal 4 issue on the nvidia system.

stultumanto · 2024-07-31 04:27:22

I booted the system off another partition with Daedalus and X starts fine. If I'm understanding the backtrace, it appears that the nouveau kernel module (nouveau_dri.so) is causing a hardware exception (SIGILL) which would make sense, as this issue started to occur after a kernel upgrade. I'm not sure if I'm reading it right, though. I would like to make a proper bug report, but I don't where it should go, as this problem involves the X server, display driver, and kernel.

Update: Apparently it wasn't the kernel after all. I rebooted with multiple kernel versions, but the X server always fails to start. I guess I'll stick with Daedalus for now. Expecting this old nvidia card to work reliably on the testing branch may be asking too much.

Last edited by stultumanto (2024-08-03 01:13:35)

stultumanto · 2024-08-27 03:27:34

So, many Excalibur updates later, and this is still occurring. I tried manually changing the driver for both video cards to modesetting, and while X no longer loaded the nouveau driver, it still crashed with a backtrace pointing to /usr/lib/x86_66-linux-gnu/dri/nouveau_dri.so. As far as I can tell, then, the nouveau kernel module is causing X to crash even when I'm not using the nouveau display driver.

There doesn't seem to be a way to get around this problem as long as I'm using this hardware. I'm OK using Daedalus for now, but I'm not sure how close Excalibur is to a stable release. If this bug survives to it, I'll have to abandon the distro entirely.

Altoid · 2024-08-27 10:58:41

Hello:

stultumanto wrote:

... old nvidia card to work reliably on the testing branch may be asking too much.

Indeed.

You do know that Devuan Excalibur is a testing suite, unlike Devuan Daedalus which is a stable suite, yes?

Getting from testing to stable takes time.
And (needless to say) a lot of work from the developers/maintainers.
ie: many, many updates and tests to run, rinse / repeat, etc.

So yes, I fully agree with you.
Expecting a testing suite to work without any issues is asking too much.

That said, my old Sun Ultra 24 workstation (ca. 2007) runs on Devuan Daedalus with a pair of old Nvidia Quadro FX580 (ca. 2009) using nouveau2/stable 2.4.114-1+b1 and picom/stable 9.1-1 (much better than compton) to feed tree old 19" monitors: a pair of SyncMaster 940n and a Dell 1914S.

~$ cat /var/log/Xorg.0.log | grep -i glamor
[    32.394] (II) Loading sub module "glamoregl"
[    32.394] (II) LoadModule: "glamoregl"
[    32.394] (II) Loading /usr/lib/xorg/modules/libglamoregl.so
[    32.404] (II) Module glamoregl: vendor="X.Org Foundation"
[    33.104] (II) modeset(0): glamor X acceleration enabled on NV96
[    33.104] (II) modeset(0): glamor initialized
[    33.293] (II) Loading sub module "glamoregl"
[    33.293] (II) LoadModule: "glamoregl"
[    33.293] (II) Loading /usr/lib/xorg/modules/libglamoregl.so
[    33.293] (II) Module glamoregl: vendor="X.Org Foundation"
[    33.303] (II) modeset(G0): glamor X acceleration enabled on NV96
[    33.303] (II) modeset(G0): glamor initialized
~$

Best,

A.

Last edited by Altoid (2024-08-27 11:00:15)

The officially official Devuan Forum!

#1 2024-07-24 21:05:41

X server - signal 4 (illegal instruction) loading glamoregl

#2 2024-07-25 07:05:09

Re: X server - signal 4 (illegal instruction) loading glamoregl

#3 2024-07-25 08:55:16

Re: X server - signal 4 (illegal instruction) loading glamoregl

#4 2024-07-31 04:27:22

Re: X server - signal 4 (illegal instruction) loading glamoregl

#5 2024-08-27 03:27:34

Re: X server - signal 4 (illegal instruction) loading glamoregl

#6 2024-08-27 10:58:41

Re: X server - signal 4 (illegal instruction) loading glamoregl

Board footer